In connection with Tech.AD Europe 2023, we spoke with Tony Karam, Lionel Bennes & Gilles Gallee from Ansys. They emphasize safety in developing autonomous vehicles, and discuss challenges and benefits of simulation tools, synthetic sensor data generation, and software to aid certification. They stress the importance of safety by design and validation in AVs’ overall safety case.
Tony Karan, can you tell us more about Ansys’ commitment to safety in developing ADAS and Autonomous Vehicles?
Tony Karan: When it comes to autonomous driving, safety is imperative. The primary engineering challenge in developing ADAS and Autonomous Vehicles is ensuring that they will operate safely under all situations. Since vehicles are likely to encounter millions of different driving situations throughout their operational life, engineering of ADAS and AVs must comprehensively address the design and validation of their software and hardware components to cover the entire spectrum of driving scenarios. OEMs and other mobility solutions providers are turning to Ansys Model-based systems engineering and Ansys simulation software to level up their driverless technologies and to achieve safety compliance faster.
Can you walk us through a customer use case where Ansys’ approach helped bring safe and reliable L3 functions to market?
Tony Karan: Notably, Mercedes-Benz is the first OEM to obtain a globally recognized Level 3 conditional automated driving certification with the S-Class sedan in May 2022. Ansys provided innovative advanced reliability methods that enable Mercedes-Benz AG to make a safety statement for Level 3 ADAS using scenario-based simulation.
Gilles Gallee, what are some significant changes needed in the validation and verification of automated driving features to meet safety standards?
Gilles Gallee: Traditionally, automakers were demonstrating performance and safety using real test driving happening late at the end of the vehicle development. In the case of autonomous technology, OEMs cannot replicate safety performance through testing alone. The methodology nodes — or body of rules, methods, and procedures — evolved with new standards describing the new methodology needed to make an adequate demonstration of safety.
For example, SOTIF (ISO PAS/21448) was specifically developed to address autonomous and semi-autonomous vehicle safety challenges faced by vehicle software developers. Simulation is the only cost effective and time efficient way — outside of an unfeasible amount of physical testing — to demonstrate AV system safety.
OEMs are relying on Ansys products to build solid and credible simulation toolchain to reach future homologation of automated driving system.
Lionel Bennes, how do you anticipate simulation at scale to help with the validation of AD system components like sensing, AI-based perception, function & controls?
Lionel Bennes: To develop a robust product that has a 1 in a million (1ppm) failure rate, one needs to conduct billions of simulation scenarios using brute force approach. For all practical purposes (w.r.t time and cost) this is not feasible.
Ansys brings a unique approach to market by combining statistics and simulation at scale. Our approach is based on 2 main pillars:
Tony Karam, how can physics-based sensor simulation help reduce field operation testing from early design stages up to V&V activities at xIL validation of ADAS/AD Systems?
Tony Karam: Edge case scenarios are the most critical to evaluate but they also come with the lowest probability of occurrence. You can drive days and nights waiting for them to happen. Therefore, simulation is only way to replicate these edge cases systematically.
Ansys’ unique capabilities to simulate all sensor modalities [i.e radar, lidar, camera], including thermal cameras, are providing the level of fidelity to produce credible simulation scenarios that mimic the real world.
Gilles Gallee, can you discuss the benefits of synthetic sensor data generation in the validation and verification process of AD systems?
Gilles Gallee: Field operation tests are cost and time prohibitive as they involve real life driving, collecting, recording, processing and storage of data. Whereas, synthetic sensor data can be generated quickly, on-demand and processed in the cloud, thereby reducing the overall costs by factor of several hundreds if not more. Synthetic data certainly brings a huge benefit, but users need to ensure that it is always generated by a trustable simulation solution that provides hi-fidelity data for all sensor modalities.
Tony Karam, how does Ansys connect safety by design and safety by V&V to drive the incremental safety case?
Tony Karam: Safety is a complex problem to solve. Both design stage and V&V stage play a critical role in establishing the overall safety case. Good thing is that Ansys has tools that address both the stages:
First is Safety by Design: Ansys medini analyze provides an end-to-end model-based safety and cybersecurity solution to accelerate autonomous system development and certification.
Second is Safety by V&V: Ansys AVxcelerate provides simulated scenarios at scale including hi fidelity sensor models.
Our customers are continuously considering this combined approach to address the safety requirements along the process with Ansys medini analyze (Safety by Design) and with AVxcelerate (Safety by Validation).
Gilles Gallee, how do you see the chasm between L2+ and L3/4 being bridged in terms of validation and verification?
Gilles Gallee: At Level 3, the onus for safety begins to shift from the driver to the OEM. Therefore, it’s even more in the OEMs interests to dedicate resources to establish a well-rounded safety case to avoid any liabilities.
Lionel Bennes, can you give us an overview of how Ansys’ AV simulation solution can aid in the certification process of AD systems?
Lionel Bennes: Mercedes-Benz’s successful foray into L3 autonomy using Ansys is really a blueprint for the market.
Similarly, we have also partnered with BMW to implement a toolchain that will bring the ability to demonstrate autonomous vehicle safety ahead of the actual safety analysis. Soon, other OEMs will undoubtedly follow the methodologies behind these achievements, especially those involving simulation.
Ansys is the primary partner at the 10th anniversary of Tech.AD Europe. How do you see Tech.AD Europe contributing to the advancement of ADAS and autonomous driving systems, and what benefits do you see in attending in this event?
Every year, TECH.AD Europe gathers a wide panel of experts sharing knowledge and experience. We are proud to be part of that community contributing to solve one of the most challenging disruptions – the development of self-driving cars!